library(groundhog)
## Loaded 'groundhog' (version:2.0.1) using R-4.2.1
## Tips and troubleshooting: https://groundhogR.com
pkgs <- c("corrr","tidyverse","lme4","lmerTest", "ggeffects", "sjPlot", "insight", "data.table", "here", "arrow", "ggpubr", "wesanderson")
library(grid)
groundhog.day <- '2022-07-25'
groundhog.library(pkgs, groundhog.day)
## Registered S3 method overwritten by 'seriation':
## method from
## reorder.hclust gclus
## [36mSuccesfully attached 'corrr_0.4.3'[0m
## Registered S3 methods overwritten by 'readr':
## method from
## as.data.frame.spec_tbl_df vroom
## as_tibble.spec_tbl_df vroom
## format.col_spec vroom
## print.col_spec vroom
## print.collector vroom
## print.date_names vroom
## print.locale vroom
## str.col_spec vroom
## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.2 ──
## ✔ ggplot2 3.3.6 ✔ purrr 0.3.4
## ✔ tibble 3.1.8 ✔ dplyr 1.0.9
## ✔ tidyr 1.2.0 ✔ stringr 1.4.0
## ✔ readr 2.1.2 ✔ forcats 0.5.1
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## [36mSuccesfully attached 'tidyverse_1.3.2'[0m
##
## Loading required package: Matrix
##
##
## Attaching package: 'Matrix'
##
##
## The following objects are masked from 'package:tidyr':
##
## expand, pack, unpack
##
##
## [36mSuccesfully attached 'lme4_1.1-30'[0m
##
##
## Attaching package: 'lmerTest'
##
##
## The following object is masked from 'package:lme4':
##
## lmer
##
##
## The following object is masked from 'package:stats':
##
## step
##
##
## [36mSuccesfully attached 'lmerTest_3.1-3'[0m
##
## [36mSuccesfully attached 'ggeffects_1.1.2'[0m
##
## [36mSuccesfully attached 'sjPlot_2.8.10'[0m
##
## [36mSuccesfully attached 'insight_0.18.0'[0m
##
##
## Attaching package: 'data.table'
##
##
## The following objects are masked from 'package:dplyr':
##
## between, first, last
##
##
## The following object is masked from 'package:purrr':
##
## transpose
##
##
## [36mSuccesfully attached 'data.table_1.14.2'[0m
##
## here() starts at /Users/jacobelder/Documents/GitHub/NormIdeol
##
## [36mSuccesfully attached 'here_1.0.1'[0m
##
##
## Attaching package: 'arrow'
##
##
## The following object is masked from 'package:utils':
##
## timestamp
##
##
## [36mSuccesfully attached 'arrow_8.0.0'[0m
##
## [36mSuccesfully attached 'ggpubr_0.4.0'[0m
##
## [36mSuccesfully attached 'wesanderson_0.3.6'[0m
here::i_am("main/Analysis/publicationPlots.Rmd")
## here() starts at /Users/jacobelder/Documents/GitHub/NormIdeol
devtools::source_url("https://raw.githubusercontent.com/JacobElder/MiscellaneousR/master/plotCommAxes.R")
## ℹ SHA-1 hash of file is "374a4de7fec345d21628a52c0ed0e4f2c389df8e"
objectName <- function(input) {
deparse(match.call()$input)
}
#study 1
longDf1 <- read_parquet(here("main","data","PFfullDf.parquet"))
distDf1 <- read_parquet(here("main","data","distanceDf.parquet"))
varDf1 <- read_parquet(here("main","data","varDf1.parquet"))
study1_ind <- read.csv(here("main","data","Study1_indDiff.csv"))
study1_ind<- study1_ind[,c(2, 4,5,6,7,8,9)]
longDf1 <- merge(longDf1, study1_ind, by = c("subID"), all.x = T)
longDf1$partyN <- as.factor(longDf1$partyN)
longDf1$Rep <- as.factor(longDf1$Rep)
contrasts(longDf1$Rep) <- contr.sum(3)
longDf1$RepN <- as.factor(longDf1$RepN)
longDf1$RepN <- relevel(longDf1$RepN,"In")
longDf1$Info <- as.factor(longDf1$Info)
contrasts(longDf1$Info) <- contr.sum(2)
longDf1$partyN <- as.factor(longDf1$partyN)
contrasts(longDf1$partyN) <- contr.sum(2)
demDf <- subset(longDf1, partyN == "Dem")
repDf <- subset(longDf1, partyN == "Rep")
longDf1$politStre <- as.factor(abs(4-longDf1$Polit))
InfoDf1 <- subset(longDf1, Info == "Info")
issueDf1 <- longDf1[!duplicated(longDf1$issues),]
distDf1$polStrength <- as.factor(abs(4-distDf1$Polit))
#study 2
longDf2 <- read_parquet(here("main","data","PFfullDf2.parquet"))
distDf2 <- read_parquet(here("main","data","distanceDf2.parquet"))
varDf2 <- read_parquet(here("main","data","varDf2.parquet"))
study2_ind <- read.csv(here("main","data","Study2_indDiff.csv"))
study2_ind<- study2_ind[,c(2:11, 13:15)]
longDf2 <- merge(longDf2, study2_ind, by = c("subID"), all.x = T)
longDf2$partyN <- as.factor(longDf2$partyN)
longDf2$Rep <- as.factor(longDf2$Rep)
longDf2$Rep <- factor(longDf2$Rep, c("Non","Rep","Dem"))
longDf2$Rep <- relevel(longDf2$Rep,"Non")
contrasts(longDf2$Rep) <- contr.sum(3)
longDf2$RepN <- as.factor(longDf2$RepN)
longDf2$RepN <- relevel(longDf2$RepN,"In")
longDf2$Info <- as.factor(longDf2$Info)
contrasts(longDf2$Info) <- contr.sum(2)
longDf2$partyN <- as.factor(longDf2$partyN)
contrasts(longDf2$partyN) <- contr.sum(2)
demDf <- subset(longDf2, partyN == "Dem")
repDf <- subset(longDf2, partyN == "Rep")
longDf2$politStre <- as.factor(abs(4-longDf2$Polit))
InfoDf2 <- subset(longDf2, Info == "Info")
issueDf2 <- longDf2[!duplicated(longDf2$issues),]
plotswd <- ("/Volumes/GoogleDrive/My Drive/Volumes/Research Project/Preference Falsification/1+2Comb/Plots/")
Info and Similarity
Study 1
m <- lmer( dist ~ corr * Info + ( corr | subID), data = distDf1, control=lmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
summary(m)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: dist ~ corr * Info + (corr | subID)
## Data: distDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## REML criterion at convergence: 7263613
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9079 -0.7748 -0.1156 0.6898 4.4950
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2388 0.4887
## corr 2.0669 1.4377 -0.66
## Residual 2.9317 1.7122
## Number of obs: 1854945, groups: subID, 384
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.32162 0.03531 381.92951 65.745 <2e-16 ***
## corr -1.69016 0.10389 382.00480 -16.269 <2e-16 ***
## InfoNon -0.03438 0.04994 381.93358 -0.689 0.4916
## corr:InfoNon 0.32378 0.14692 382.00867 2.204 0.0281 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) corr InfoNn
## corr -0.655
## InfoNon -0.707 0.463
## corr:InfoNn 0.463 -0.707 -0.655
p <- ggpredict(m, c("corr","Info"))
Sim.Info.S1 <-ggplot(p, aes(x, predicted)) + geom_line(aes(linetype=group, color=group)) + geom_ribbon(aes(ymin=conf.low, ymax=conf.high, fill=group), alpha=0.15) + scale_linetype_discrete(labels = c("Group Norms","No Group Norms")) + scale_color_manual(labels = c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + scale_fill_manual(
labels=c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + theme(
legend.position = c(.6, .65),
legend.justification = c("left", "bottom"),
legend.box.just = "left",
legend.margin = margin(6, 6, 6, 6)
) + theme(axis.text=element_text(size=12),
axis.title=element_text(size=12,face="bold")) + theme(legend.text = element_text(size=12)) + theme(panel.border = element_rect(colour = "black", fill = NA, size =1)) + theme(legend.title = element_blank()) + theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black")) +
xlab("Issue Similarity") + ylab("Rating Dissimilarity")
Sim.Info.S1
Raw data
ggpredict(m, c("corr","Info"))%>% plot(add.data=T) + labs(title="Raw Data",x="Issue Similarity",y="Evaluation Similarity")
Study 2
m <- lmer( dist ~ corr * Info + ( corr | subID), data = distDf2, control=lmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
summary(m)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: dist ~ corr * Info + (corr | subID)
## Data: distDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## REML criterion at convergence: 17357713
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0270 -0.7620 -0.1053 0.7012 4.1755
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 101.5 10.08
## corr 742.2 27.24 -0.67
## Residual 735.7 27.12
## Number of obs: 1838503, groups: subID, 380
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 38.94716 0.73737 377.98588 52.819 <2e-16 ***
## corr -31.58797 1.99453 377.95846 -15.837 <2e-16 ***
## InfoNon 0.08693 1.03467 377.99769 0.084 0.9331
## corr:InfoNon 4.90766 2.79870 377.97216 1.754 0.0803 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) corr InfoNn
## corr -0.672
## InfoNon -0.713 0.479
## corr:InfoNn 0.479 -0.713 -0.672
p <- ggpredict(m, c("corr","Info"))
Sim.Info.S2 <-ggplot(p, aes(x, predicted)) + geom_line(aes(linetype=group, color=group)) + geom_ribbon(aes(ymin=conf.low, ymax=conf.high, fill=group), alpha=0.15) + scale_linetype_discrete(labels = c("Group Norms","No Group Norms")) + scale_color_manual(labels = c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + scale_fill_manual(
labels=c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + theme(
legend.position = c(.6, .65),
legend.justification = c("left", "bottom"),
legend.box.just = "left",
legend.margin = margin(6, 6, 6, 6)
) + theme(axis.text=element_text(size=12),
axis.title=element_text(size=12,face="bold")) + theme(legend.text = element_text(size=12)) + theme(panel.border = element_rect(colour = "black", fill = NA, size =1)) + theme(legend.title = element_blank()) + theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black")) +
xlab("Issue Similarity") + ylab("Rating Dissimilarity") + scale_x_continuous(breaks=seq(-1,1,1))
Sim.Info.S2
ggpredict(m, c("corr","Info"))%>% plot(add.data=T) + labs(title="Raw Data",x="Issue Similarity",y="Evaluation Similarity")
Sim.Info.Comb <- plotCommAxes(Sim.Info.S1, Sim.Info.S2, xaxis = "Issue Similarity", yaxis = "Rating Dissimilarity")
Sim.Info.Comb
ggsave(paste0(plotswd,objectName(Sim.Info.Comb),".png"), width = 9, height = 6, dpi=500, units="in")
ggsave(paste0(plotswd,objectName(Sim.Info.Comb),".tiff"), width = 9, height = 6, dpi=500, units="in")
Affective Polarization and Similarity
Study 1
m <- lmer( dist ~ corr * affPol + ( corr | subID), data = distDf1, control=lmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
summary(m)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: dist ~ corr * affPol + (corr | subID)
## Data: distDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## REML criterion at convergence: 7245702
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9052 -0.7751 -0.1160 0.6905 4.4931
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2155 0.4642
## corr 1.9184 1.3850 -0.62
## Residual 2.9340 1.7129
## Number of obs: 1849995, groups: subID, 383
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.070e+00 4.341e-02 3.810e+02 47.674 < 2e-16 ***
## corr -8.919e-01 1.295e-01 3.811e+02 -6.886 2.38e-11 ***
## affPol 4.212e-03 6.489e-04 3.810e+02 6.491 2.66e-10 ***
## corr:affPol -1.143e-02 1.936e-03 3.810e+02 -5.902 7.96e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) corr affPol
## corr -0.619
## affPol -0.837 0.518
## corr:affPol 0.518 -0.837 -0.619
p <- ggpredict(m, c("corr","affPol"))
Sim.AffPol.S1 <-ggplot(p, aes(x, predicted)) + geom_line(aes(linetype=group, color=group)) + geom_ribbon(aes(ymin=conf.low, ymax=conf.high, fill=group), alpha=0.15) + scale_linetype_discrete(labels = c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol.")) + scale_color_manual(labels = c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol."), values = wes_palette("Darjeeling1")) + scale_fill_manual(
labels=c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol."), values = wes_palette("Darjeeling1")) + theme(
legend.position = c(.6, .65),
legend.justification = c("left", "bottom"),
legend.box.just = "left",
legend.margin = margin(6, 6, 6, 6)
) + theme(axis.text=element_text(size=12),
axis.title=element_text(size=12,face="bold")) + theme(legend.text = element_text(size=12)) + theme(panel.border = element_rect(colour = "black", fill = NA, size =1)) + theme(legend.title = element_blank()) + theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black")) +
xlab("Issue Similarity") + ylab("Rating Dissimilarity")
Sim.AffPol.S1
ggpredict(m, c("corr","affPol"))%>% plot(add.data=T) + labs(title="Raw Data",x="Issue Similarity",y="Evaluation Similarity")
Study 2
m <- lmer( dist ~ corr * affPol + ( corr | subID), data = distDf2, control=lmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
summary(m)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: dist ~ corr * affPol + (corr | subID)
## Data: distDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## REML criterion at convergence: 17357678
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0269 -0.7620 -0.1054 0.7012 4.1752
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 87.43 9.35
## corr 686.52 26.20 -0.63
## Residual 735.74 27.12
## Number of obs: 1838503, groups: subID, 380
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 33.43278 0.85958 377.98743 38.894 < 2e-16 ***
## corr -17.45695 2.40945 377.96187 -7.245 2.45e-12 ***
## affPol 0.10659 0.01367 377.99089 7.796 6.27e-14 ***
## corr:affPol -0.22319 0.03833 377.96296 -5.823 1.23e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) corr affPol
## corr -0.632
## affPol -0.829 0.524
## corr:affPol 0.524 -0.829 -0.632
p <- ggpredict(m, c("corr","affPol"))
Sim.AffPol.S2 <-ggplot(p, aes(x, predicted)) + geom_line(aes(linetype=group, color=group)) + geom_ribbon(aes(ymin=conf.low, ymax=conf.high, fill=group), alpha=0.15) + scale_linetype_discrete(labels = c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol.")) + scale_color_manual(labels = c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol."), values = wes_palette("Darjeeling1")) + scale_fill_manual(
labels=c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol."), values = wes_palette("Darjeeling1")) + theme(
legend.position = c(.6, .65),
legend.justification = c("left", "bottom"),
legend.box.just = "left",
legend.margin = margin(6, 6, 6, 6)
) + theme(axis.text=element_text(size=12),
axis.title=element_text(size=12,face="bold")) + theme(legend.text = element_text(size=12)) + theme(panel.border = element_rect(colour = "black", fill = NA, size =1)) + theme(legend.title = element_blank()) + theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black")) +
xlab("Issue Similarity") + ylab("Rating Dissimilarity") + scale_x_continuous(breaks=seq(-1,1,1))
Sim.AffPol.S2
ggpredict(m, c("corr","affPol"))%>% plot(add.data=T) + labs(title="Raw Data",x="Issue Similarity",y="Evaluation Similarity")
Sim.AffPol.Comb <- plotCommAxes(Sim.AffPol.S1, Sim.AffPol.S2, xaxis = "Issue Similarity", yaxis = "Rating Dissimilarity")
Sim.AffPol.Comb
ggsave(paste0(plotswd,objectName(Sim.AffPol.Comb),".png"), width = 9, height = 6, dpi=500, units="in")
ggsave(paste0(plotswd,objectName(Sim.AffPol.Comb),".tiff"), width = 9, height = 6, dpi=500, units="in")
Sim.Comb <- ggarrange(Sim.AffPol.Comb, Sim.Info.Comb, nrow = 2, ncol = 1)
Sim.Comb
ggsave(paste0(plotswd,objectName(Sim.Comb),".png"), width = 10, height = 9, dpi=500, units="in")
ggsave(paste0(plotswd,objectName(Sim.Comb),".tiff"), width = 10, height = 9, dpi=500, units="in")
Ideological Polarization
Info
Study 1
m <- lmer( eval ~ inMinOut*Info + ( inMinOut | subID) + (1 | issues), data = longDf1, control=lmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
summary(m)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: eval ~ inMinOut * Info + (inMinOut | subID) + (1 | issues)
## Data: longDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## REML criterion at convergence: 141384.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4206 -0.6319 0.0958 0.6610 4.3618
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.1246 0.3531
## inMinOut 2.0690 1.4384 -0.45
## issues (Intercept) 1.0658 1.0324
## Residual 2.3036 1.5178
## Number of obs: 37917, groups: subID, 384; issues, 100
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.71874 0.10526 104.82038 44.828 < 2e-16 ***
## inMinOut 2.15317 0.07651 380.61536 28.141 < 2e-16 ***
## Info1 -0.01271 0.01973 380.43750 -0.644 0.51981
## inMinOut:Info1 0.20690 0.07651 380.62042 2.704 0.00716 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) inMnOt Info1
## inMinOut -0.073
## Info1 0.000 0.000
## inMnOt:Inf1 -0.001 0.000 -0.394
p <- ggpredict(m, c("inMinOut","Info"))
Norm.Info.S1 <-ggplot(p, aes(x, predicted)) + geom_line(aes(linetype=group, color=group)) + geom_ribbon(aes(ymin=conf.low, ymax=conf.high, fill=group), alpha=0.15) + scale_linetype_discrete(labels = c("Group Norms","No Group Norms")) + scale_color_manual(labels = c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + scale_fill_manual(
labels=c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + theme(
legend.position = c(.6, .20),
legend.justification = c("left", "bottom"),
legend.box.just = "left",
legend.margin = margin(6, 6, 6, 6)
) + theme(axis.text=element_text(size=12),
axis.title=element_text(size=12,face="bold")) + theme(legend.text = element_text(size=12)) + theme(panel.border = element_rect(colour = "black", fill = NA, size =1)) + theme(legend.title = element_blank()) + theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black")) +
xlab("Inter-Party Favorability") + ylab("Approval Rating") + scale_y_continuous(breaks=seq(1,7,2)) + scale_x_continuous(breaks=seq(-1,1,1))
Norm.Info.S1
ggpredict(m, c("inMinOut","Info"))%>% plot(add.data=T) + labs(title="Raw Data",x="Inter-Party Favorability",y="Evaluation")
Study 2
m <- lmer( eval ~ inMinOut*Info + ( inMinOut | subID) + (1 | issues), data = longDf2, control=lmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
summary(m)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: eval ~ inMinOut * Info + (inMinOut | subID) + (1 | issues)
## Data: longDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## REML criterion at convergence: 350609.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2083 -0.6302 0.0562 0.6536 3.7512
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 44.94 6.704
## inMinOut 538.14 23.198 -0.33
## issues (Intercept) 350.28 18.716
## Residual 627.06 25.041
## Number of obs: 37549, groups: subID, 380; issues, 100
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 56.7959 1.9097 104.2910 29.740 < 2e-16 ***
## inMinOut 30.2353 1.2478 376.9426 24.230 < 2e-16 ***
## Info1 -0.7493 0.3688 378.1852 -2.032 0.04285 *
## inMinOut:Info1 3.9351 1.2501 377.3046 3.148 0.00178 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) inMnOt Info1
## inMinOut -0.055
## Info1 0.003 -0.006
## inMnOt:Inf1 -0.004 0.014 -0.290
p <- ggpredict(m, c("inMinOut","Info"))
Norm.Info.S2 <-ggplot(p, aes(x, predicted)) + geom_line(aes(linetype=group, color=group)) + geom_ribbon(aes(ymin=conf.low, ymax=conf.high, fill=group), alpha=0.15) + scale_linetype_discrete(labels = c("Group Norms","No Group Norms")) + scale_color_manual(labels = c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + scale_fill_manual(
labels=c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + theme(
legend.position = c(.6, .20),
legend.justification = c("left", "bottom"),
legend.box.just = "left",
legend.margin = margin(6, 6, 6, 6)
) + theme(axis.text=element_text(size=12),
axis.title=element_text(size=12,face="bold")) + theme(legend.text = element_text(size=12)) + theme(panel.border = element_rect(colour = "black", fill = NA, size =1)) + theme(legend.title = element_blank()) + theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black")) +
xlab("Issue Similarity") + ylab("Rating Dissimilarity") + scale_x_continuous(breaks=seq(-1,1,1))
Norm.Info.S2
ggpredict(m, c("inMinOut","Info"))%>% plot(add.data=T) + labs(title="Raw Data",x="Inter-Party Favorability",y="Evaluation")
Combined
Norm.Info.Comb <- plotCommAxes(Norm.Info.S1, Norm.Info.S2, xaxis = "Inter-Party Favorability", yaxis = "Approval Rating")
Norm.Info.Comb
ggsave(paste0(plotswd,objectName(Norm.Info.Comb),".png"), width = 9, height = 6, dpi=500, units="in")
ggsave(paste0(plotswd,objectName(Norm.Info.Comb),".tiff"), width = 9, height = 6, dpi=500, units="in")
Affective Polarization
Study 1
m <- lmer( eval ~ inMinOut*affPol + ( inMinOut | subID) + (1 | issues), data = longDf1, control=lmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
summary(m)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: eval ~ inMinOut * affPol + (inMinOut | subID) + (1 | issues)
## Data: longDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## REML criterion at convergence: 140939.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4298 -0.6336 0.0957 0.6610 4.3573
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.1237 0.3517
## inMinOut 1.5461 1.2434 -0.47
## issues (Intercept) 1.0434 1.0215
## Residual 2.3050 1.5182
## Number of obs: 37817, groups: subID, 383; issues, 100
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.755e+00 1.084e-01 1.230e+02 43.859 < 2e-16 ***
## inMinOut 9.867e-01 1.237e-01 3.829e+02 7.978 1.74e-14 ***
## affPol -8.214e-04 5.379e-04 3.797e+02 -1.527 0.128
## inMinOut:affPol 2.089e-02 1.855e-03 3.838e+02 11.263 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) inMnOt affPol
## inMinOut -0.124
## affPol -0.278 0.327
## inMnOt:ffPl 0.101 -0.840 -0.392
p <- ggpredict(m, c("inMinOut","affPol"))
Norm.affPol.S1 <-ggplot(p, aes(x, predicted)) + geom_line(aes(linetype=group, color=group)) + geom_ribbon(aes(ymin=conf.low, ymax=conf.high, fill=group), alpha=0.15) + scale_linetype_discrete(labels = c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol.")) + scale_color_manual(labels = c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol."), values = wes_palette("Darjeeling1")) + scale_fill_manual(
labels=c("Low Aff. Pol.","Medium Aff. Pol.", "High Aff. Pol."), values = wes_palette("Darjeeling1")) + theme(
legend.position = c(.6, .20),
legend.justification = c("left", "bottom"),
legend.box.just = "left",
legend.margin = margin(6, 6, 6, 6)
) + theme(axis.text=element_text(size=12),
axis.title=element_text(size=12,face="bold")) + theme(legend.text = element_text(size=12)) + theme(panel.border = element_rect(colour = "black", fill = NA, size =1)) + theme(legend.title = element_blank()) + theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black")) +
xlab("Inter-Party Favorability") + ylab("Approval Rating") + scale_x_continuous(breaks=seq(-1,1,1)) + scale_y_continuous(breaks=seq(1,7,2))
Norm.affPol.S1
ggpredict(m, c("inMinOut","affPol"))%>% plot(add.data=T) + labs(title="Raw Data",x="Inter-Party Favorability",y="Evaluation")
Study 2
m <- lmer( eval ~ inMinOut*affPol + ( inMinOut | subID) + (1 | issues), data = longDf2, control=lmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
summary(m)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: eval ~ inMinOut * affPol + (inMinOut | subID) + (1 | issues)
## Data: longDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## REML criterion at convergence: 350506.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2341 -0.6306 0.0558 0.6539 3.6990
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 43.1 6.565
## inMinOut 380.2 19.500 -0.23
## issues (Intercept) 338.0 18.385
## Residual 627.1 25.041
## Number of obs: 37549, groups: subID, 380; issues, 100
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 58.88778 1.95105 122.21960 30.183 < 2e-16 ***
## inMinOut 10.41790 1.93358 378.28310 5.388 1.26e-07 ***
## affPol -0.04348 0.01031 377.97543 -4.217 3.10e-05 ***
## inMinOut:affPol 0.37797 0.03082 378.48039 12.262 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) inMnOt affPol
## inMinOut -0.057
## affPol -0.276 0.156
## inMnOt:ffPl 0.045 -0.834 -0.191
p <- ggpredict(m, c("inMinOut","affPol"))
Norm.affPol.S2 <-ggplot(p, aes(x, predicted)) + geom_line(aes(linetype=group, color=group)) + geom_ribbon(aes(ymin=conf.low, ymax=conf.high, fill=group), alpha=0.15) + scale_linetype_discrete(labels = c("Group Norms","No Group Norms")) + scale_color_manual(labels = c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + scale_fill_manual(
labels=c("Group Norms","No Group Norms"), values = wes_palette("Darjeeling1")) + theme(
legend.position = c(.6, .20),
legend.justification = c("left", "bottom"),
legend.box.just = "left",
legend.margin = margin(6, 6, 6, 6)
) + theme(axis.text=element_text(size=12),
axis.title=element_text(size=12,face="bold")) + theme(legend.text = element_text(size=12)) + theme(panel.border = element_rect(colour = "black", fill = NA, size =1)) + theme(legend.title = element_blank()) + theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
panel.background = element_blank(), axis.line = element_line(colour = "black")) +
xlab("Inter-Party Favorability") + ylab("Approval Rating") + scale_x_continuous(breaks=seq(-1,1,1))
Norm.affPol.S2
ggpredict(m, c("inMinOut","affPol"))%>% plot(add.data=T) + labs(title="Raw Data",x="Inter-Party Favorability",y="Evaluation")
Combined
Norm.affPol.Comb <- plotCommAxes(Norm.affPol.S1, Norm.affPol.S2, xaxis = "Inter-Party Favorability", yaxis = "Approval Rating")
Norm.affPol.Comb
ggsave(paste0(plotswd,objectName(Norm.affPol.Comb),".png"), width = 9, height = 6, dpi=500, units="in")
ggsave(paste0(plotswd,objectName(Norm.affPol.Comb),".tiff"), width = 9, height = 6, dpi=500, units="in")
Norm.Comb <- ggarrange(Norm.affPol.Comb, Norm.Info.Comb, nrow = 2, ncol = 1)
Norm.Comb
ggsave(paste0(plotswd,objectName(Norm.Comb),".png"), width = 10, height = 9, dpi=500, units="in")
ggsave(paste0(plotswd,objectName(Norm.Comb),".tiff"), width = 10, height = 9, dpi=500, units="in")